Vehicle driving scene judgment method and device, vehicle and storage medium

By constructing three-dimensional view and bird's eye view, the behavior information of road interaction objects is determined, which solves the problem of low accuracy in driving scenario judgment caused by drivers' reliance on subjective judgment, and improves driving safety.

CN119068458APending Publication Date: 2024-12-03ZHEJIANG GEELY HLDG GRP CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411122795.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

In the prior art, drivers need to rely on subjective judgments to determine the behavior of road interaction objects around the vehicle when driving a vehicle, resulting in low accuracy in determining driving scenarios and affecting driving safety.

Method used

By collecting spatial information around the target vehicle, building a three-dimensional view and a bird's-eye view, determining the behavioral information of the road interaction object, and determining the driving scenario based on this information to improve the judgment accuracy.

Benefits of technology

It improves the accuracy of vehicle driving scenario judgment, enhances driving safety, and reduces the dependence on drivers' subjective judgment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119068458A_ABST
    Figure CN119068458A_ABST
Patent Text Reader

Abstract

The invention discloses a vehicle driving scene judgment method and device, a vehicle and a storage medium, and relates to the technical field of automatic driving, and the method comprises the steps: collecting space information in a preset detection range around a target vehicle, and constructing a three-dimensional view and a bird's-eye view based on the space information; first behavior information of the first road interaction objects in the three-dimensional view is determined, and the first behavior information represents behavior intentions of the first road interaction objects; second behavior information of each second road interaction object in the aerial view is determined, and the second behavior information represents behavior intentions of the second road interaction objects; and determining a driving scene where the target vehicle is located based on the first behavior information and the second behavior information. According to the invention, the accuracy of determining the vehicle driving scene can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a method, device, vehicle and storage medium for determining a vehicle driving scene. Background Art

[0002] At present, when driving a vehicle, the driver not only needs to control the vehicle's driving direction and speed, but also needs to pay attention to other road interaction objects around the vehicle (such as vehicles, pedestrians and obstacles), and judge the behavior and intention of other road interaction objects based on their own experience and intuition, so as to make the next driving decision accordingly. However, this method relies too much on the driver's subjective judgment and lacks the support of objective data. There is a problem of low accuracy in driving scene judgment, which affects the safety of the driver's driving.

[0003] Therefore, how to improve the accuracy of judging vehicle driving scenarios is a problem that needs to be solved urgently. Summary of the invention

[0004] The main purpose of this application is to provide a method, device, vehicle and storage medium for determining a vehicle driving scene, aiming to improve the accuracy of determining the vehicle driving scene.

[0005] To achieve the above object, the present application provides a method for determining a vehicle driving scene, the method comprising:

[0006] Collect spatial information within a preset detection range around the target vehicle, and construct a three-dimensional view and a bird's-eye view based on the spatial information;

[0007] Determining first behavior information of each first road interaction object in the three-dimensional view, wherein the first behavior information represents a behavior intention of the first road interaction object;

[0008] Determining second behavior information of each second road interaction object in the bird's-eye view, wherein the second behavior information represents a behavior intention of the second road interaction object;

[0009] The driving scene of the target vehicle is determined based on each piece of the first behavior information and each piece of the second behavior information.

[0010] In one embodiment, the bird's-eye view takes the position of the front-view camera of the target vehicle in the spatial information as the coordinate origin and takes the road interaction object farthest from the coordinate origin within the preset detection range as the end point.

[0011] In one embodiment, the step of determining the first behavior information of each first road interaction object in the three-dimensional view includes:

[0012] Determine a three-dimensional detection frame of each first road interactive object in the three-dimensional view, wherein the three-dimensional detection frame represents a position of the first road interactive object in the three-dimensional view;

[0013] Tracking each of the three-dimensional detection frames to obtain respective historical trajectory information of each of the first road interaction objects;

[0014] The first behavior information of each of the first road interaction objects is determined based on each of the historical trajectory information.

[0015] In one embodiment, the step of determining the second behavior information of each second road interaction object in the bird's-eye view includes:

[0016] Determining a bird's-eye view feature set for each second road interaction object in the bird's-eye view;

[0017] Determine a query vector for each of the second road interactive objects based on the bird's-eye view and each of the bird's-eye feature sets, wherein the query vector represents a visual feature of the second road interactive object;

[0018] Each of the query vectors and preset prompt words is input into the first target model to obtain the second behavior information of each of the second road interaction objects, wherein the first target model is trained using the sample behavior information and preset prompt words of each third road interaction object in the sample bird's-eye view of the vehicle as model input data and the real driving scene in which the vehicle is located as the model training label, wherein the sample bird's-eye view is constructed based on the spatial information within the preset detection range around the vehicle, and the preset prompt words are used to guide the output form of the second behavior information.

[0019] In one embodiment, the step of determining the query vector of each of the second road interaction objects based on the bird's-eye view and each of the bird's-eye feature sets comprises:

[0020] The bird's-eye view and each of the bird's-eye feature sets are input into a second target model to obtain a query vector for each of the second road interaction objects, wherein the second target model is trained using a sample bird's-eye view of the vehicle and a sample bird's-eye feature set of each third road interaction object in the sample bird's-eye view as model input data, and using a feature vector constructed based on the real visual features of each of the third road interaction objects as a model training label.

[0021] In one embodiment, the step of determining the bird's-eye view feature set of each second road interaction object in the bird's-eye view comprises:

[0022] Matching each of the first road interaction objects and each of the second road interaction objects to obtain each object group, wherein the object group includes the first road interaction object and the second road interaction object used to represent the same road interaction object;

[0023] A bird's-eye view feature set of each of the second road interaction objects is determined based on each of the object groups.

[0024] In one embodiment, the step of determining the bird's-eye view feature set of each of the second road interaction objects based on each of the object groups comprises:

[0025] Determine the location information of the first road interactive object at different historical moments based on the historical trajectory information of the first road interactive object in the target object group, wherein the target object group is any object group in each of the object groups;

[0026] respectively determining the bird's-eye view spaces that match the respective position information in the bird's-eye view, sampling the respective bird's-eye view spaces, and obtaining the respective bird's-eye view features of the respective bird's-eye view spaces;

[0027] A bird's-eye view feature set of the second road interaction object in the target object group is determined based on each of the bird's-eye view features.

[0028] In one embodiment, the step of determining the driving scene of the target vehicle based on each of the first behavior information and each of the second behavior information includes:

[0029] Determining common behavior information in each of the first behavior information and each of the second behavior information;

[0030] Determine the preset priorities corresponding to each of the shared behavior information, and determine the target behavior information with the highest priority among the shared behavior information as the driving scene where the target vehicle is located.

[0031] In addition, to achieve the above-mentioned purpose, the present application also provides a vehicle driving scene determination device, the vehicle driving scene determination device comprising:

[0032] A construction module, used to collect spatial information within a preset detection range around the target vehicle, and construct a three-dimensional view and a bird's-eye view based on the spatial information;

[0033] A first determination module, configured to determine first behavior information of each first road interaction object in the three-dimensional view, wherein the first behavior information represents a behavior intention of the first road interaction object;

[0034] A second determination module is used to determine second behavior information of each second road interaction object in the bird's-eye view, wherein the second behavior information represents a behavior intention of the second road interaction object;

[0035] A determination module is used to determine the driving scene of the target vehicle based on each of the first behavior information and each of the second behavior information.

[0036] In addition, to achieve the above-mentioned purpose, the present application also provides a storage medium, which is a computer-readable storage medium, and the computer-readable storage medium stores a program for implementing a method for determining a vehicle driving scene, and the program for implementing a method for determining a vehicle driving scene is executed by a processor to implement the steps of the method for determining a vehicle driving scene as described above.

[0037] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method for determining a vehicle driving scene as described above.

[0038] The present application provides a method for determining a vehicle driving scene. The present application collects spatial information within a preset detection range around a target vehicle, constructs a three-dimensional view and a bird's-eye view based on the collected spatial information, determines first behavior information of each first road interaction object in the three-dimensional view, and determines second behavior information of each second road interaction object in the bird's-eye view, wherein the first behavior information represents the behavior intention of the first road interaction object, and the second behavior information represents the behavior intention of the second road interaction object. Finally, the driving scene where the target vehicle is located is determined based on each first behavior information and each second behavior information.

[0039] In this way, compared with the traditional method that requires the driver to manually monitor the vehicle's surrounding environment to make driving decisions, the present application collects spatial information around the target vehicle, constructs a three-dimensional view and a bird's-eye view of the vehicle's surrounding environment based on the spatial information, that is, constructs view information in two dimensions, and then determines the first behavior information of the first road interaction object in the three-dimensional view and the second behavior information of the second road interaction object in the bird's-eye view respectively, and then combines the first behavior information used to characterize the behavior intention of the first road interaction object and the second behavior information used to characterize the behavior intention of the second road interaction object to determine the current driving scene of the target vehicle, that is, models the spatial information around the target vehicle that affects the vehicle's driving decision, and determines the behavior intentions of the road interaction objects in the modeled three-dimensional view and the bird's-eye view respectively, and then determines the driving scene of the target vehicle based on the determined first behavior information and second behavior information, thereby improving the accuracy of vehicle driving scene determination and thereby improving driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0042] Figure 1 This is a flow chart of a first embodiment of a method for determining a vehicle driving scene of the present application;

[0043] Figure 2 A schematic diagram of a spatial information processing flow involved in an embodiment of a method for determining a vehicle driving scene of the present application;

[0044] Figure 3 A schematic diagram of constructing a BEV view involved in an embodiment of a method for determining a vehicle driving scene of the present application;

[0045] Figure 4 This is a schematic diagram of object matching involved in an embodiment of a method for determining a vehicle driving scene of the present application;

[0046] Figure 5 A schematic diagram of bird's-eye view feature extraction involved in an embodiment of a method for determining a vehicle driving scene of the present application;

[0047] Figure 6 A schematic diagram of a driving scene determination process involved in an embodiment of a method for determining a vehicle driving scene of the present application;

[0048] Figure 7 This is a schematic diagram of the module structure of the vehicle driving scene determination device of the present application;

[0049] Figure 8 A schematic diagram of the vehicle structure of the hardware operating environment involved in the method for determining a vehicle driving scenario in an embodiment of the present application.

[0050] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0051] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0052] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0053] The main solution of the present application is: to collect spatial information within a preset detection range around the target vehicle, and to construct a three-dimensional view and a bird's-eye view based on the spatial information; to determine the first behavior information of each first road interaction object in the three-dimensional view, wherein the first behavior information represents the behavior intention of the first road interaction object; to determine the second behavior information of each second road interaction object in the bird's-eye view, wherein the second behavior information represents the behavior intention of the second road interaction object; and to determine the driving scene where the target vehicle is located based on each of the first behavior information and each of the second behavior information.

[0054] At present, when driving a vehicle, the driver not only needs to control the vehicle's driving direction and speed, but also needs to pay attention to other road interaction objects around the vehicle (such as vehicles, pedestrians and obstacles), and judge the behavior and intention of other road interaction objects based on their own experience and intuition, so as to make the next driving decision accordingly. However, this method relies too much on the driver's subjective judgment and lacks the support of objective data. There is a problem of low accuracy in driving scene judgment, which affects the safety of the driver's driving.

[0055] In this way, compared with the traditional method that requires the driver to manually monitor the vehicle's surrounding environment to make driving decisions, the present application collects spatial information around the target vehicle, constructs a three-dimensional view and a bird's-eye view of the vehicle's surrounding environment based on the spatial information, that is, constructs view information in two dimensions, and then determines the first behavior information of the first road interaction object in the three-dimensional view and the second behavior information of the second road interaction object in the bird's-eye view respectively, and then combines the first behavior information used to characterize the behavior intention of the first road interaction object and the second behavior information used to characterize the behavior intention of the second road interaction object to determine the current driving scene of the target vehicle, that is, models the spatial information around the target vehicle that affects the vehicle's driving decision, and determines the behavior intentions of the road interaction objects in the modeled three-dimensional view and the bird's-eye view respectively, and then determines the driving scene of the target vehicle based on the determined first behavior information and second behavior information, thereby improving the accuracy of vehicle driving scene determination and thereby improving driving safety.

[0056] In addition, it should be noted that the behavior prediction of road interactors in autonomous driving can serve as a reference for the next action planning of the self-driving car, enabling the self-driving car to make safe driving decisions, which is a key technology in autonomous driving. The basis of behavior prediction is the self-driving car's perception of the surrounding environment, and the perception of key information such as roads, vehicles, pedestrians, obstacles, etc., especially for moving objects in space, it is necessary to analyze their past trajectories for prediction. Compared with the perspective of the self-driving car, BEV (Bird's Eye View) can observe global information and observe moving objects in static space, which can better predict the movement direction and movement behavior of road interactors, and ultimately define the current driving scene. The process of behavior prediction in this application is as follows: data collection, using a single camera to collect information around the vehicle body; BEV space modeling, projecting the spatial information collected by a single camera into the BEV space; road perception, in the process of autonomous driving, real-time perception of road information is crucial, and this application perceives road information, obstacle information, vehicle information, and pedestrian information in real time; tracking, tracking the perceived objects; trajectory generation, generating the trajectory of the interactor in the BEV space; behavior prediction and description, the action prediction set in this application, including collision with the vehicle in front, going straight, inserting the vehicle in front, turning left, turning right, going straight with the lateral vehicle, going straight with the oncoming vehicle, overtaking the vehicle in front, slowing down / stationary of the vehicle in front, etc. The large model performs behavior judgment and intention analysis based on the trajectory generated by each tracking object. In this way, this application can generate a judgment on the behavior of the interactor based on the trajectory of the road interactor generated by the model, and finally match it with several set actions.

[0057] It should be noted that the execution subject of the method in each embodiment of the vehicle driving scene determination method of the present application can be a driving scene determination system, or a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or a vehicle capable of realizing the above functions, etc., and this embodiment does not specifically limit this. The following takes the driving scene determination system as an example to illustrate this embodiment and the following embodiments.

[0058] Based on this, the present application proposes a method for determining a vehicle driving scene in a first embodiment, please refer to Figure 1 The method for determining the vehicle driving scene includes steps S10 to S40:

[0059] Step S10, collecting spatial information within a preset detection range around the target vehicle, and constructing a three-dimensional view and a bird's-eye view based on the spatial information;

[0060] It should be noted that the target vehicle is the vehicle currently driven by the user. The above-mentioned preset detection range at least includes the range where the vehicle interaction objects that can affect the driving decision of the target vehicle are located. The specific size of the detection range is determined by the performance of the detection equipment. The three-dimensional view is the 3D (3-Dimensional) view, and the bird's-eye view is the BEV view.

[0061] The spatial information within the preset detection range around the target vehicle is collected, and a three-dimensional view and a bird's-eye view of the target vehicle's surrounding environment are constructed based on the spatial information.

[0062] In a specific implementation, the step S10 may include:

[0063] Step S101, collecting spatial information within a preset detection range around the target vehicle through the forward-looking camera of the target vehicle;

[0064] The spatial information within a preset detection range around the target vehicle is collected by the forward-looking camera of the target vehicle. That is, it should be understood that the spatial information is the image collected by the forward-looking camera, and the preset detection range is the maximum range that the forward-looking camera can capture.

[0065] Step S102 : projecting the spatial information into a three-dimensional space to obtain a three-dimensional view of the spatial information, and projecting the spatial information into a bird's-eye view space to obtain a bird's-eye view of the spatial information.

[0066] The spatial information collected by the front-view camera is projected into a three-dimensional space to obtain a three-dimensional view of the spatial information, and the spatial information is projected into a bird's-eye view space to obtain a bird's-eye view of the spatial information.

[0067] For example, Figure 2The figure shows a schematic diagram of the spatial information processing flow. The front-view camera of the target vehicle collects images of normal traffic flow within a preset detection range around the target vehicle in real time. The front-view camera can record videos or set a collection cycle to collect images once in each collection cycle. Then, each frame of image (i.e., front view) collected by the front-view camera is detected, tracked, and mapped. Specifically, for the extracted image frames, Vi t (Vision Transformer) + FPN (Feature Pyramid Network) and other methods are used to build a feature pyramid for the image frame. Then, the DETR3D (3D Object Detection from Multi-view Images via 3D-to-2D Queries, a three-dimensional object detection algorithm based on multi-view images) detection algorithm is used to detect the road interaction objects in each image frame, and the detected road interaction objects are projected into 3D space according to the internal and external parameters of the front-view camera to obtain a 3D view (i.e., the above-mentioned three-dimensional view).

[0068] In this embodiment, the bird's-eye view takes the position of the front-view camera of the target vehicle in the spatial information as the coordinate origin, and takes the road interaction object farthest from the coordinate origin within the preset detection range as the end point.

[0069] It should be noted that if Figure 3 The figure shows a schematic diagram of BEV view construction. The traditional BEV view construction method usually sets up a fixed-size detection range, such as horizontal [-50m, 50m], vertical [-50m, 50m], and height [-5m, 3m]. The vehicle is located at the center of the view, and a 100*100*256 BEV Query is established based on this. The embodiment of the present application uses the vehicle's front-view camera as the coordinate origin and the farthest detected target as the end point to establish the BEV range, with a height of [-1m, 5.4m]. Its BEVQuery feature is still 100*100*256. That is, in traditional BEV-related technologies, the detection is performed to establish a scene for each frame of a multi-perspective image, in which the vehicle is located at the exact center of the scene. The scene is centered on the vehicle, ignoring the movement of the vehicle, and only observing the relative movement of other road interaction objects with respect to the vehicle. This is not conducive to the observation of the motion trajectory of each road interaction object in absolute coordinates. The embodiment of the present application uses the position of the vehicle in the first frame as the origin to establish an absolute coordinate system, so that the motion trajectories of the vehicle and other vehicles can be clearly seen.

[0070] In the image frame captured by the front-view camera, the position of the front-view camera is taken as the coordinate origin, and the road interaction object farthest from the coordinate origin among the captured road interaction objects is taken as the end point to determine the bird's-eye view space, so as to project the spatial information into the bird's-eye view space to obtain a bird's-eye view.

[0071] Step S20, determining first behavior information of each first road interaction object in the three-dimensional view, wherein the first behavior information represents the behavior intention of the first road interaction object;

[0072] It should be noted that the behavior information of the road interaction object represents the position of the road interaction object relative to the target vehicle, the type of the road interaction object and the driving behavior of the road interaction object (i.e., behavior intention). The types of road interaction objects include but are not limited to vehicles, pedestrians and obstacles, etc. The driving behavior of the road interaction object includes but is not limited to stationary and crossing, etc. For example, the driving behavior includes but is not limited to collision with the vehicle ahead, cutting in sideways and pedestrians crossing ahead, etc.

[0073] Behavior information (hereinafter referred to as first behavior information for distinction) of each road interaction object (hereinafter referred to as first road interaction object for distinction) in the three-dimensional view is determined.

[0074] In a specific implementation, the step S20 may include:

[0075] Step S201, determining a three-dimensional detection frame of each first road interactive object in the three-dimensional view, wherein the three-dimensional detection frame represents a position of the first road interactive object in the three-dimensional view;

[0076] Step S202, tracking each of the three-dimensional detection frames to obtain respective historical trajectory information of each of the first road interaction objects;

[0077] Step S203: determining first behavior information of each of the first road interaction objects based on each of the historical trajectory information.

[0078] It should be noted that the 3D detection frame of the first road interactive object represents the position of the first road interactive object in the 3D view. The historical track information of the first road interactive object includes the position information of the first road interactive object at different historical moments.

[0079] A three-dimensional detection frame of each first road interaction object is determined in the three-dimensional view, and each three-dimensional detection frame is tracked to obtain historical trajectory information of each first road interaction object, and then first behavior information of each first road interaction object is determined based on the historical trajectory information.

[0080] Exemplarily, for the detected 3D frame (i.e., three-dimensional detection frame), the MoMA-M3T algorithm is used to match and track each of the first road interaction objects. Scene judgment is performed based on the historical trajectory information of each first road interaction object and the preset judgment rules. Taking the stationary vehicle in front as an example, the behavior of the target vehicle is judged by detecting the neighboring vehicles in front, speed changes, and position changes. If the speed of the neighboring target in front of multiple consecutive frames is small and the position is almost unchanged, the vehicle in front is judged to be stationary. The near end is within 30 meters from the vehicle, the far end is 30 meters away, and the detection angle range is [0,30]. The behavior of each vehicle is classified by the behavior judgment algorithm, and then the most concerned scene is detected.

[0081] Step S30, determining second behavior information of each second road interaction object in the bird's-eye view, wherein the second behavior information represents the behavior intention of the second road interaction object;

[0082] Determine the behavior information (hereinafter referred to as second behavior information for distinction) of each road interaction object (hereinafter referred to as second road interaction object for distinction) in the bird's-eye view, wherein the second behavior information represents the behavior intention of the second road interaction object.

[0083] In this embodiment, the step S30 may include:

[0084] Step S301, determining a bird's-eye view feature set of each second road interaction object in the bird's-eye view; it should be noted that the bird's-eye view feature set of the second road interaction object represents the position of the second road interaction object at different historical moments in the bird's-eye view and the type of the second road interaction object.

[0085] Step S302, determining a query vector for each of the second road interaction objects based on the bird's-eye view and each of the bird's-eye view feature sets, wherein the query vector represents a visual feature of the second road interaction object;

[0086] Based on the bird's-eye view and the bird's-eye feature set of each second road interaction object in the bird's-eye view, a query vector of each second road interaction object is determined, wherein it should be understood that the query vector represents the visual features of the second road interaction object in the bird's-eye view.

[0087] In this embodiment, the step S302 may include:

[0088] Step S3021, input the bird's-eye view and each of the bird's-eye feature sets into a second target model to obtain a query vector for each of the second road interaction objects, wherein the second target model is trained using a sample bird's-eye view of the vehicle and a sample bird's-eye feature set of each third road interaction object in the sample bird's-eye view as model input data, and using a feature vector constructed based on the real visual features of each of the third road interaction objects as a model training label.

[0089] It should be noted that the vehicle's bird's-eye view (hereinafter referred to as the sample bird's-eye view for distinction) and the bird's-eye feature sets (hereinafter referred to as the sample bird's-eye feature sets for distinction) of each road interaction object in the sample bird's-eye view (hereinafter referred to as the third road interaction object for distinction) are used as model input data in advance, and the feature vectors constructed based on the real visual features of each third road interaction object are used as model training labels to train the initial LLAMA model to obtain a trained LLAMA model (hereinafter referred to as the second target model for distinction). Among them, the sample bird's-eye view is constructed based on the spatial information within the preset detection range around the vehicle.

[0090] Input the bird's-eye view and the bird's-eye feature set of each second road interaction object in the bird's-eye view into the second target model to obtain a query vector for characterizing the visual features of each second road interaction object. Step S303: Input each query vector and preset prompt word into the first target model to obtain the second behavior information of each second road interaction object, wherein the first target model is trained using the sample behavior information and preset prompt words of each third road interaction object in the sample bird's-eye view of the vehicle as model input data and the real driving scene where the vehicle is located as the model training label, wherein the sample bird's-eye view is constructed based on the spatial information within the preset detection range around the vehicle, and the preset prompt word is used to guide the output form of the second behavior information. It should be noted that the initial Qformer model is trained in advance with the sample bird's-eye view of the vehicle (and the sample bird's-eye feature set of each third road interaction object in the sample bird's-eye view as the model input data, and the real driving scene in which the vehicle is located is used as the model training label, to obtain the trained Qformer model (hereinafter referred to as the first target model for distinction). In addition, the prompt words for guiding the output form of the second behavior information output by the first target model are pre-set, that is, the above-mentioned preset prompt words. Exemplarily, the preset prompt words can be "vehicle turning around on the left", "pedestrians crossing" and "rear vehicle overtaking", etc. It should be understood that the first target model outputs the second behavior information based on the expression form of the preset prompt words.

[0091] Exemplarily, the bird's-eye view and the bird's-eye view feature set of each second road interaction object in the bird's-eye view are input into the second target model. Visual language matching is performed through the trained Qformer (i.e., the second target model mentioned above), and then the query vector output by the second target model and the pre-set prompt word are input into the LLAMA model (i.e., the first target model mentioned above), and according to the prompt word, the driving scene where the target vehicle is currently located and the behavior information of each second road interaction object (hereinafter referred to as the second behavior information for distinction) are obtained, thereby enhancing the interpretability of the model.

[0092] Step S40: determining a driving scene of the target vehicle based on each of the first behavior information and each of the second behavior information.

[0093] Based on the first behavior information of each first road interaction object in the three-dimensional view and the second behavior information of each second road interaction object in the bird's-eye view, a driving scene where the target vehicle is located is determined.

[0094] In this way, compared with the traditional method that requires the driver to manually monitor the vehicle's surrounding environment to make driving decisions, the embodiment of the present application collects spatial information around the target vehicle, constructs a three-dimensional view and a bird's-eye view of the vehicle's surrounding environment based on the spatial information, that is, constructs two-dimensional view information, and then determines the first behavior information of the first road interaction object in the three-dimensional view and the second behavior information of the second road interaction object in the bird's-eye view respectively, and then combines the first behavior information used to characterize the behavior intention of the first road interaction object and the second behavior information used to characterize the behavior intention of the second road interaction object to determine the current driving scene of the target vehicle, that is, models the spatial information around the target vehicle that affects the vehicle's driving decision, and determines the behavior intentions of the road interaction objects in the modeled three-dimensional view and the bird's-eye view respectively, and then determines the driving scene of the target vehicle based on the determined first behavior information and second behavior information, thereby improving the accuracy of vehicle driving scene determination and thereby improving driving safety.

[0095] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. On this basis, the step S301 may include:

[0096] Step A10, matching each of the first road interaction objects and each of the second road interaction objects to obtain each object group, wherein the object group includes the first road interaction object and the second road interaction object used to represent the same road interaction object;

[0097] Each first road interaction object in the three-dimensional view and each second road interaction object in the bird's-eye view are matched to obtain object groups, wherein one object group includes a first road interaction object and a second road interaction object that are matched consistently, and the first road interaction object and the second road interaction object that represent the same road interaction object are considered to be a matched consistently object group.

[0098] For example, Figure 4 The figure shows a schematic diagram of object matching. For each first road interaction object, it has its own 3D coordinates. For the learned BEVQuery, a detection head is added to perform BEV detection under the BEV perspective. For each second road interaction object detected, it is matched with the first road interaction object in the three-dimensional view. The matching method uses the Hungarian algorithm to ensure that BEV learns the 3D spatial position of each target.

[0099] Step A20: determining a bird's-eye view feature set of each of the second road interaction objects based on each of the object groups.

[0100] Based on the associated first road interaction objects and second road interaction objects in each object group, a bird's-eye view feature set of each second road interaction object is determined.

[0101] In this embodiment, the step A20 may include:

[0102] Step A201, determining the position information of the first road interactive object at different historical moments based on the historical trajectory information of the first road interactive object in the target object group, wherein the target object group is any object group in each of the object groups;

[0103] Any object group in each object group is referred to as a target object group for distinction. For the target object group, based on the historical track information of the first road interactive object in the target object group, the position information of the first road interactive object at different historical moments is determined.

[0104] Step A202, respectively determining the bird's-eye view spaces that match the respective position information in the bird's-eye view, sampling the respective bird's-eye view spaces, and obtaining the respective bird's-eye view features of the respective bird's-eye view spaces;

[0105] A bird's-eye view space matching each position information is determined in the bird's-eye view, and each bird's-eye view space is sampled to obtain a bird's-eye view feature of each bird's-eye view space, wherein a bird's-eye view feature of the second road interaction object represents the position of the second road interaction object at a certain historical moment in the bird's-eye view and the type of the second road interaction object.

[0106] Step A203: determining a bird's-eye view feature set of the second road interaction object in the target object group based on each of the bird's-eye view features.

[0107] The obtained bird's-eye view features constitute a target object group to summarize the bird's-eye view feature set of the second road interactive object. Similarly, a bird's-eye view feature set of each second road interactive object in each object group is obtained.

[0108] For example, Figure 5 The figure shows a schematic diagram of bird's-eye view feature extraction. For each first road interactive object, the DETR3D detection algorithm is used to determine the detection frame of each first road interactive object in the three-dimensional view, and then the MoMA-M3T algorithm is used to track each detection frame to obtain the historical trajectory information of each first road interactive object. That is, in the three-dimensional view, its trajectory is determinable, so the time-series 3D coordinates of the tracked target can be obtained, that is, the historical trajectory information of each first road interactive object. However, in the BEV view, its trajectory is uncertain, and BEVQuery does not show which blocks represent its detection targets. Therefore, the sampling point method is used to sample the BEV space that matches the position of any first road interactive object, and 9 neighboring points are sampled for each BEV space. Finally, the bird's-eye view features of each BEV space are obtained to form a bird's-eye view feature set.

[0109] In this way, the embodiment of the present application matches each first road interaction object in the three-dimensional view with each second road interaction object in the bird's-eye view, so that the target model learns the historical trajectory information of each second road interaction object, thereby improving the accuracy of the target model's prediction results, that is, improving the accuracy of the driving scene judgment.

[0110] Based on the first embodiment and / or the second embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those in the first embodiment can be referred to the above description, and will not be described in detail later. On this basis, the step S40 may include:

[0111] Step S401, determining common behavior information in each of the first behavior information and each of the second behavior information;

[0112] Step S402: determining preset priorities corresponding to the respective shared behavior information, and determining the target behavior information with the highest priority among the shared behavior information as the driving scene where the target vehicle is located.

[0113] It should be noted that the degree of influence of various behavior information of road interaction objects on the driving decision of the target vehicle, that is, the priority of the behavior information, is pre-set. The behavior information with a higher priority has a greater influence on the driving decision of the target vehicle than the behavior information with a lower priority. Specifically, a scene sorting retrieval table is pre-established, and the list sorts the most concerned scenes, such as a car accident ahead, a ghost head, a car cutting in from the left, a car cutting in from the right, etc.

[0114] After obtaining the first behavior information of each first road interaction object determined based on the three-dimensional view and the second behavior information of each second road interaction object determined based on the bird's-eye view, the common behavior information in each first behavior information and each second behavior information is determined, wherein it should be noted that the common behavior information belongs to both the first behavior information and the second behavior information. Then, the preset priority corresponding to each common behavior information is determined, and the behavior information with the highest priority in each common behavior information (hereinafter referred to as the target behavior information for distinction) is determined as the driving scene where the target vehicle is located.

[0115] For example, Figure 6 The figure shows a schematic diagram of the driving scene determination process, in which a three-dimensional view and a bird's-eye view are constructed based on the collected spatial information around the target vehicle; the first behavior information of each first road interaction object is determined based on the historical trajectory information of each first road interaction object in the three-dimensional view and the preset determination rule; at the same time, the bird's-eye feature set corresponding to each second road interaction object in each bird's-eye view is input into a pre-trained Qformer model to obtain a query vector of each second road interaction object, and then each query vector and the preset prompt word are input into a pre-trained LLAMA model to obtain a global-level scene description and the second behavior information of the object set, wherein it should be noted that the global-level scene description is the behavior information that has the greatest impact on the driving decision of the target vehicle among the second behavior information determined by the target model; then the common behavior information among each first behavior information and each second behavior information is determined; in the preset scene sorting retrieval table, the behavior information with the highest priority among the common behavior information is determined as the driving scene where the target vehicle is currently located.

[0116] In this way, the embodiment of the present application combines the first behavior information of each first road interaction object determined based on the three-dimensional view and the second behavior information of each second road interaction object determined based on the bird's-eye view, selects the common behavior information in the behavior information determined by the two behavior determination methods to participate in the priority comparison, and finally determines that the common behavior information with the highest priority is the driving scene where the target vehicle is currently located, thereby avoiding the problem of inaccurate final driving scene caused by misjudgment of a single behavior determination method, thereby improving the accuracy of driving scene determination.

[0117] The present application also provides a vehicle driving scene determination device, please refer to Figure 7 , the vehicle driving scene determination device comprises:

[0118] A construction module 10 is used to collect spatial information within a preset detection range around the target vehicle, and construct a three-dimensional view and a bird's-eye view based on the spatial information;

[0119] A first determination module 20, configured to determine first behavior information of each first road interaction object in the three-dimensional view, wherein the first behavior information represents a behavior intention of the first road interaction object;

[0120] A second determination module 30 is configured to determine second behavior information of each second road interaction object in the bird's-eye view, wherein the second behavior information represents a behavior intention of the second road interaction object;

[0121] The determination module 40 is used to determine the driving scene of the target vehicle based on each of the first behavior information and each of the second behavior information.

[0122] Optionally, the bird's-eye view takes the position of the target vehicle's front-view camera in the spatial information as a coordinate origin, and takes the road interaction object farthest from the coordinate origin within the preset detection range as an end point.

[0123] Optionally, the first determining module 20 is further configured to:

[0124] Determine a three-dimensional detection frame of each first road interactive object in the three-dimensional view, wherein the three-dimensional detection frame represents a position of the first road interactive object in the three-dimensional view;

[0125] Tracking each of the three-dimensional detection frames to obtain respective historical trajectory information of each of the first road interaction objects;

[0126] The first behavior information of each of the first road interaction objects is determined based on each of the historical trajectory information.

[0127] Optionally, the second determining module 30 is further configured to:

[0128] Determining a bird's-eye view feature set for each second road interaction object in the bird's-eye view;

[0129] Determine a query vector for each of the second road interactive objects based on the bird's-eye view and each of the bird's-eye feature sets, wherein the query vector represents a visual feature of the second road interactive object;

[0130] Each of the query vectors and preset prompt words is input into the first target model to obtain the second behavior information of each of the second road interaction objects, wherein the first target model is trained using the sample behavior information and preset prompt words of each third road interaction object in the sample bird's-eye view of the vehicle as model input data and the real driving scene in which the vehicle is located as the model training label, wherein the sample bird's-eye view is constructed based on the spatial information within the preset detection range around the vehicle, and the preset prompt words are used to guide the output form of the second behavior information.

[0131] Optionally, the second determining module 30 is further configured to:

[0132] The bird's-eye view and each of the bird's-eye feature sets are input into a second target model to obtain a query vector for each of the second road interaction objects, wherein the second target model is trained using a sample bird's-eye view of the vehicle and a sample bird's-eye feature set of each third road interaction object in the sample bird's-eye view as model input data, and using a feature vector constructed based on the real visual features of each of the third road interaction objects as a model training label.

[0133] Optionally, the second determining module 30 is further configured to:

[0134] Matching each of the first road interaction objects and each of the second road interaction objects to obtain each object group, wherein the object group includes the first road interaction object and the second road interaction object used to represent the same road interaction object;

[0135] A bird's-eye view feature set of each of the second road interaction objects is determined based on each of the object groups.

[0136] Optionally, the second determining module 30 is further configured to:

[0137] Determine the location information of the first road interactive object at different historical moments based on the historical trajectory information of the first road interactive object in the target object group, wherein the target object group is any object group in each of the object groups;

[0138] respectively determining bird's-eye view spaces in the bird's-eye view that match the respective position information, sampling the respective bird's-eye view spaces, and obtaining respective bird's-eye view features of the respective bird's-eye view spaces;

[0139] A bird's-eye view feature set of the second road interaction object in the target object group is determined based on each of the bird's-eye view features.

[0140] Optionally, the determining module 40 is further configured to:

[0141] Determining common behavior information in each of the first behavior information and each of the second behavior information;

[0142] Determine the preset priorities corresponding to each of the shared behavior information, and determine the target behavior information with the highest priority among the shared behavior information as the driving scene where the target vehicle is located.

[0143] The vehicle driving scene determination device provided in the embodiment of the present application adopts the vehicle driving scene determination method in the above embodiment, which can solve the technical problem of how to improve the accuracy of vehicle driving scene determination. Compared with the prior art, the beneficial effects of the vehicle driving scene determination device provided in the embodiment of the present application are the same as the beneficial effects of the vehicle driving scene determination method provided in the above embodiment, and the other technical features in the vehicle driving scene determination device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0144] The present application provides a vehicle, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method for determining the vehicle driving scene in the above-mentioned embodiment one.

[0145] Reference below Figure 8 , which shows a schematic diagram of a structure of a vehicle suitable for implementing the embodiment of the present application. The vehicle in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 8 The vehicle shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0146] like Figure 8As shown, the vehicle may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which may perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read On ly Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. Various programs and data required for vehicle operation are also stored in RAM1004. The processing device 1001, ROM1002, and RAM1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems may be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 may allow the vehicle to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a vehicle with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have instead.

[0147] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0148] The vehicle provided by the present application adopts the method for determining the vehicle driving scene in the above embodiment, which can solve the technical problem of how to improve the accuracy of determining the vehicle driving scene. Compared with the prior art, the beneficial effects of the vehicle provided by the present application are the same as the beneficial effects of the method for determining the vehicle driving scene provided by the above embodiment, and the other technical features in the vehicle are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0149] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0150] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0151] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the method for determining the vehicle driving scene in the above-mentioned embodiment.

[0152] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM: Erasable Programmab le Read On ly Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read On ly Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0153] The computer-readable storage medium may be included in the vehicle, or may exist independently without being installed in the vehicle.

[0154] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the vehicle, the vehicle: collects spatial information within a preset detection range around the target vehicle, and constructs a three-dimensional view and a bird's-eye view based on the spatial information; determines the historical trajectory information of each first road interaction object in the three-dimensional view, and determines the first behavior information of each first road interaction object based on each historical trajectory information; determines the bird's-eye feature set of each second road interaction object in the bird's-eye view, and inputs the bird's-eye view and each bird's-eye feature set into the target model to obtain the second behavior information of each second road interaction object; determines the driving scene where the target vehicle is located based on each first behavior information and each second behavior information.

[0155] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0156] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0157] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0158] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned vehicle driving scene determination method, and can solve the technical problem of how to improve the accuracy of vehicle driving scene determination. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the vehicle driving scene determination method provided by the above-mentioned embodiment, and will not be repeated here.

[0159] An embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method for determining a vehicle driving scenario as described above.

[0160] The computer program product provided by the present application can improve the accuracy of determining the vehicle driving scene. Compared with the prior art, the beneficial effects of the computer program product provided by the embodiment of the present application are the same as the beneficial effects of the method for determining the vehicle driving scene provided by the above embodiment, which will not be repeated here.

[0161] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent processing scope of the present application.

Claims

1. A method for determining a vehicle driving scene, characterized in that: The method for determining the vehicle driving scene includes: Collect spatial information within a preset detection range around the target vehicle, and construct a three-dimensional view and a bird's-eye view based on the spatial information; Determining first behavior information of each first road interaction object in the three-dimensional view, wherein the first behavior information represents a behavior intention of the first road interaction object; Determining second behavior information of each second road interaction object in the bird's-eye view, wherein the second behavior information represents a behavior intention of the second road interaction object; The driving scene of the target vehicle is determined based on each piece of the first behavior information and each piece of the second behavior information.

2. The method according to claim 1, characterized in that The bird's-eye view takes the position of the target vehicle's front-view camera in the spatial information as the coordinate origin and takes the road interaction object farthest from the coordinate origin within the preset detection range as the end point.

3. The method according to claim 1, characterized in that The step of determining first behavior information of each first road interaction object in the three-dimensional view includes: Determine a three-dimensional detection frame of each first road interactive object in the three-dimensional view, wherein the three-dimensional detection frame represents a position of the first road interactive object in the three-dimensional view; Tracking each of the three-dimensional detection frames to obtain respective historical trajectory information of each of the first road interaction objects; The first behavior information of each of the first road interaction objects is determined based on each of the historical trajectory information.

4. The method according to claim 1, characterized in that The step of determining the second behavior information of each second road interaction object in the bird's-eye view includes: Determining a bird's-eye view feature set for each second road interaction object in the bird's-eye view; Determine a query vector for each of the second road interactive objects based on the bird's-eye view and each of the bird's-eye feature sets, wherein the query vector represents a visual feature of the second road interactive object; Each of the query vectors and preset prompt words is input into the first target model to obtain the second behavior information of each of the second road interaction objects, wherein the first target model is trained using the sample behavior information and preset prompt words of each third road interaction object in the sample bird's-eye view of the vehicle as model input data and the real driving scene in which the vehicle is located as the model training label, wherein the sample bird's-eye view is constructed based on the spatial information within the preset detection range around the vehicle, and the preset prompt words are used to guide the output form of the second behavior information.

5. The method according to claim 4, characterized in that The step of determining the query vector of each of the second road interaction objects based on the bird's-eye view and each of the bird's-eye view feature sets comprises: The bird's-eye view and each of the bird's-eye feature sets are input into a second target model to obtain a query vector for each of the second road interaction objects, wherein the second target model is trained using a sample bird's-eye view of the vehicle and a sample bird's-eye feature set of each third road interaction object in the sample bird's-eye view as model input data, and using a feature vector constructed based on the real visual features of each of the third road interaction objects as a model training label.

6. The method according to claim 4, characterized in that The step of determining the bird's-eye view feature set of each second road interaction object in the bird's-eye view comprises: Matching each of the first road interaction objects and each of the second road interaction objects to obtain each object group, wherein the object group includes the first road interaction object and the second road interaction object used to represent the same road interaction object; A bird's-eye view feature set of each of the second road interaction objects is determined based on each of the object groups.

7. The method according to claim 6, characterized in that The step of determining the bird's-eye view feature set of each of the second road interaction objects based on each of the object groups comprises: Determine the location information of the first road interactive object at different historical moments based on the historical trajectory information of the first road interactive object in the target object group, wherein the target object group is any object group in each of the object groups; respectively determining bird's-eye view spaces in the bird's-eye view that match the respective position information, sampling the respective bird's-eye view spaces, and obtaining respective bird's-eye view features of the respective bird's-eye view spaces; A bird's-eye view feature set of the second road interaction object in the target object group is determined based on each of the bird's-eye view features.

8. The method according to any one of claims 1 to 7, characterized in that The step of determining the driving scene where the target vehicle is located based on each of the first behavior information and each of the second behavior information includes: Determining common behavior information in each of the first behavior information and each of the second behavior information; Determine the preset priorities corresponding to each of the shared behavior information, and determine the target behavior information with the highest priority among the shared behavior information as the driving scene where the target vehicle is located.

9. A vehicle driving scene determination device, characterized in that: The vehicle driving scene determination device comprises: A construction module, used to collect spatial information within a preset detection range around the target vehicle, and construct a three-dimensional view and a bird's-eye view based on the spatial information; A first determination module, configured to determine first behavior information of each first road interaction object in the three-dimensional view, wherein the first behavior information represents a behavior intention of the first road interaction object; A second determination module is used to determine second behavior information of each second road interaction object in the bird's-eye view, wherein the second behavior information represents a behavior intention of the second road interaction object; A determination module is used to determine the driving scene of the target vehicle based on each of the first behavior information and each of the second behavior information.

10. A vehicle, characterized in that: The vehicle comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the method for determining a vehicle driving scenario according to any one of claims 1 to 8.

11. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for determining a vehicle driving scene according to any one of claims 1 to 8 are implemented.

12. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the method for determining a vehicle driving scene according to any one of claims 1 to 8 are implemented.